Most of your points are valid - but not all. So, going a-nitpicking...
fork() has its good, bad and ugly things. To be expected, as it comes from a pre-threading era. So any comparison with CreateThread() is unfair (you might as well compare it with pthread_create() ). And it's different enough in functionality from CreateProcess() (again, you wouldn't be comparing fork() with exec*(), would you?) Note that my point is not 'good/bad', only 'not the same kettle of fish'
pipes are about loose integration. Again, you're making an unfair comparison - different tools for different purposes.
I take you meant using VBScript for pulling an address off Outlook's Address Book. If you use a different mail client, you're back to the 'which address book' can of worms. Preaching ease of use is easy when your environment is as predictable as a default Windows one (whether this predictability is a strong point or not is a separate issue)
remote X... well, it was mostly addressed already. One thing to note is compatibility - you can get remote X sessions between vastly different servers, both in terms of vendor and in terms of version (btw, how ancient of an X server were those slow Unix stations running?). Again, it's a thing with goods, bads and uglies, but so is rdesktop.
umm... NTFS... hairy topic. For one thing the 'fast' claim is a fake; the honest one is 'a lot faster than the FAT32 piece of crap'. It's fast, alright... out of the box. Try again after one month of heavy use, especially at more than 75% full. Defragmentation is still a regular task in Windows. Granted, it has lots of features - but you're mixing things again, stuff like transparent compression and encryption are fs hooks (reparse points?) - these particular ones tightly integrated in the OS and sitting on top of NTFS. Just like the likes of WinFS will. In particular, encryption is not as good an option as it looks - 128bit asymmetric keys were good enough for win2k, but are rather weak nowadays.
NTFS had journalling capabilities since Win2k - hardly the first mainstream journaled fs, but quite before ext3 was anything useable, indeed.
ahh... the good old copy/paste issue in X. Yes, it does not always work. No, it does not always work in Windows, either - but people conveniently forget that Windows backward compatibility for this thing is not entirely golden. With X, I can do copy/paste against Motif apps, or Athena Widgets ones (meaning really old legacy apps; yes, some big tin servers do use those). Try pasting to a win3.x app from a winXP-targeted one - if the clipboard format is not recognized, tough luck. Granted, for win32 it's one of the strong points, but amazingly enough, X has most of the framework in place, it's just not standardized. You should be able to write 2 generic X apps that could copy-paste pretty much anything that can be copy/pasted in Windows - negotiate your selection type and how you pass info around. Too bad it's not widely used - most of the X copy/paste mechanism complaints could actually go away.
Finally, and not really a main point - why on earth would you expect Alt+Tab to switch windows in Linux??? just because it's the standard Windows behavior? That, my friend, IS narrow-minded thinking.
I's a bit surprised your post stood rather unchallenged so long. But I do agree with your main point - the GP poster picked really poor examples:-))
2': Apple zealots who are fanatically dedicated to MacOSX. 5. Regular '* is dying', 'In Soviet Russia', et alt. trolls.
Still, your argument holds - the Mac fans won't bettray their faith anymore than the Linux or *BSD ones (less, actually, if history has any meaning); and the troll mentality hints they're already windows users, so there's nothing to gain there.
I guess the only target would be the platform agnostic developers - but they'll not just swallow the MS ads about ease of use/devel and TCO hook, sink and line as MS would want. Oh well, maybe the MS PR knows more than I do about the benefits of add-spamming your audience.
What's the science icon doing on top for this article? It's only Apple's PR, whether the clusters get awarded to some science, defense or giant panda groups it's not really relevant. Put the science icon on some news about what the winners did with the clusters, when those will come out.
Unfortunately you're not entirely correct. Most people considering Linux vs. Windows from an end-user perspective would do something on the lines of Linux stands for 'RedHat Linux' which is equated to its UI, hence Gnome. They'll say "this is how Linux looks and behaves" while looking at RH's version of the Gnome UI.
And that is because people coming from Windows don't have a high-enough resolution radar to tell the kernel apart from the UI and the producer of the distro. They look at the whole package, see the UI and call it Linux - because this is what they did with Windows all the time! Heck, even average OSX users don't single out the interface from the rest of the system!
One can only hope they'll go far enough with Linux to be able to look at the UI and call it by its proper name, be that Gnome, KDE or anything else.
So if it's a half-truth, it's not worth it, right? ok, let's get rid of some of the 'wrong' halves (the ones making fun):
separate widows suck for file browsing - if you don't think so, you haven't had to work with deep directory structures and may God preserve your innocence forever. It was even in Windows 3.x (remember that one? all about separate windows, it was). In Win95+/NT+ you can at least tell Explorer to use a single window from a menu option. If Nautilius forces it on unknowing users without an easy opt-out, then it sucks.
very few people will care where exactly the window for a particular folder will open - it's nice to have sometimes, but if you're constantly working with more than about 4-5 file browsers open the positions won't mean much; add multiple desktops in the picture and you've got a mess in your hands - what did I have in the upper left corner of desktop 3?
So here - 2 examples of design decisions that only appealed to very few users, while obeying some arcane UI design 'rules'. Remember, UI design is an empirical science, so if more than 50% of the users don't like some setting it's the theory that is more likely to be wrong, not the users. There's no provision against making experiments in the UI, but it's quite rude to force them on people who will have close to no chance of figuring out how to change the default annoying behavior to something they can live with. And exactly this kind of 'we know better even if 95% of the users disagree' ego trip attitude was one of the big UI critiques of proprietary systems, so it would be really sad if F/OSS ends up taking the same route.
ok, so I'm illiterate - can't tell different/. id's. sorry about putting words in your mouth. The GGP poster said that part about treating symptoms of obesity, which is funny and sad. Overeating is not an illness, it's a symptom. Obesity is a symptom, too much food and too little exercise are symptoms... it all is about how far up the cause and effect chain you feel comfortable going - and what you want to call illness.
Did you read my post actually? I never said obesity is an hereditary illness. It is an ill state of the body, but that's not all there is - there's an ill state of mind coupled with it. By thinning out one gets a chance to break the circle and do something about it. It's quite hard (though not impossible) to make a fat person - and I mean a really fat, round one - to change his/her ways. Your advice works for losing some 40lb (the belly type), but try telling that to someone who is 150lb+ overweight. Yeah, right. Too hard - that's why surgery is so frequent, if you can afford it. Mind you, I'm not saying this (if it ever makes it out of the lab) will actually cure everyone - but for many people it could be a chance.
You, on the other hand, were talking about treating obesity symptoms (2 posts up). All that does is hook you up on medication - it cures nothing. And, if obesity is the cause for those symptoms, what would happen if this condition goes away?
(side note: here's another losing weight advice: cook your own meals. More sating than fast-food waste, so you'll eat less and use more of it. Just in case you got the impression that I was disagreeing with your post entirely. It's just that I don't think there's an one size fits all cure for obesity)
If you look for a more accurate story (New Scientist? did/. fall so low already?), you'll see they're well aware of potential problems. From this yahoo story:
When fat mice were injected with the new "fat-zapper" every day for a month, they all slimmed down to normal weight with no visible side-effects, the researchers reported in the June issue of Nature Medicine.
But they stressed the experiment is still in the very early stages and it affects a function found in virtually all cells -- meaning it has a high potential for serious side-effects.
"I am trying to un-hype this," said Dr. Wadih Arap of the M.D. Anderson Cancer Center in Houston, who led the research.
You'd have to combine it with treatment for cholesterol (and other obesity symptoms).
Do you really believe that? oh well, I guess you do, judging from the rest of the post. How sad. That's probably a clue about a medical system where often enough symptoms would get treated instead of the actual disease. Here's a hint: treat the illness and the symptoms will go away (at least, when possible - and in this case the illness is what they're aiming at).
thanks for the info! to my defense, I was plannyng to read that book... but 'a bit' later (when I'll be sixty or so). I guess I have some of his other exquisite writings higher up on my lousy to do list right now ^_^
heh! too bad the mods decided to clue the clueless and mark you up as funny - a +5 insightful would have spawned a funny thread of replies to a funny post ^_^
Given the sheer number of packages included I doubt RH would have the manpower to review every single one for stolen code. They're packaging F/OSS so it would not really be a big difference if some package authors were to commit to CVS directly in terms of responsability (and it would be for fedora, not RHEL, so the responsability should be lower). In practice setting this up is usually not that easy. And the lawyers have to be consulted anyway.
The issue seems to be RH has been dragging their feet doing this. (see the 'test code, find bugs' theme). ^_^
hmm... given that the previous application was submitted some 4 years before this gives rise to an interesting perspective... say you file for some crap now and after 2-3 years snatch some new idea in the field and just amend your previous filing (ok, that's a bit overblown, but not entirely impossible). Would that make your application take precedence to any other one since the original filing date is earlier?
err... no. pair formation is wavelength dependent - it's a threshold process. And yes, ionization (the spark) depends on absorbtion - the easier you can break electrons free the better. It's a quantum transition involved so it depends on power density (through field strength, but not wavelength here) and then on wavelength (you need to overlap the laser line with the absorbtion line of the transition).
actually, the laser is probably not focused at all - and you'd have to be talking dynamic focusing in IR and handling the full laser power (near the actual output). Plus, at that power density you'll start losing energy by air ionization long before you reach the minimum beam size. Then absorbtion skyrockets and your beam power goes the way of the dodo.
If you tune to the particular laser it's not that bad - narrow-band IR filters are actually easy. The problem is heat dissipation, but you'd already have to handle that just for air friction. Te funny thing is, if you get the thermal shield to handle enough friction heat, you might just get rid of the laser due to plasma reflectivity around the missile.
The amount of power to handle isn't 100% of the output either - the beam has diffraction (decent over km ranges), the atmosphere has absorbtion (can be fairly high in IR) and so on.
Their first '93 patent refers to the '87 application (not patent) as 'now abandoned'. There's no patent issued to Dana Khoyi prior to 1993 - 5,206,951 and that one was filed in '91. So any pre-'91 prior art should be ok.
as the AC said, you have to know first there's a patent in the field. This is not trivial even for older patents. Then, for any new idea one would need to go through not only awarded patents, but pending applications also. And I'm talking obscure patents filed overseas, too. Then figure out whether any vaguely related ones can be infringing. How many lawyers per programmer would that require? any way you put it, all this does is skyrocket the R&D costs of any new application.
The patent system as it is was ok for a moderate research pace with very few players in a given field. The way things are now, you have no guarantee you don't start a completely new project and by the time you're done it will be plastered with partially covering patents held by others. It might be useful to have some of the research stages published just to be able to defend against litigation companies.
well, the first patent at least is just a spec - there's no implementation attached (the filing has only 55 pages, including figures). Moreover, they use generic examples (folders, spreadsheets) about how to use the spec, not exact details (app spreadsheet101 does this and that using ObjectManager105). The bulk of the filing are function names and descriptions, field sizes and such. These are always part of the interface spec, including CORBA.
and no, it's not Kodak that developed it - they just got the patents of a dying software company ^_^
Looking over the first patent it appears to be a library issue, not a compiler one. Unless I read it totally wrong, this could have even CORBA/COM as prior art. For example:
An object based data processing system including an extensible set of object types and a corresponding set of "object managers" wherein each object manager is a program for operating with the data stored in a corresponding type of object. The object managers in general support at least a standard set of operations. Looks a lot like interfaces 'operating on' implementations to me.
Any program can effect performance of these standard operations on objects of any type by making an "invocation" request. In response to an invocation request, object management services (which are available to all object managers) identifies and invokes an object manager that is suitable for performing the requested operation on the specified type of data. Can you spell factory class?
Data interchange services are provided for communicating data between objects of different types, using a set of standard data interchange formats. err... marshalling?
Slashdot Mirror
Finally, and not really a main point - why on earth would you expect Alt+Tab to switch windows in Linux??? just because it's the standard Windows behavior? That, my friend, IS narrow-minded thinking.
I's a bit surprised your post stood rather unchallenged so long. But I do agree with your main point - the GP poster picked really poor examples
you're missing a few types:
2': Apple zealots who are fanatically dedicated to MacOSX.
5. Regular '* is dying', 'In Soviet Russia', et alt. trolls.
Still, your argument holds - the Mac fans won't bettray their faith anymore than the Linux or *BSD ones (less, actually, if history has any meaning); and the troll mentality hints they're already windows users, so there's nothing to gain there.
I guess the only target would be the platform agnostic developers - but they'll not just swallow the MS ads about ease of use/devel and TCO hook, sink and line as MS would want. Oh well, maybe the MS PR knows more than I do about the benefits of add-spamming your audience.
What's the science icon doing on top for this article? It's only Apple's PR, whether the clusters get awarded to some science, defense or giant panda groups it's not really relevant. Put the science icon on some news about what the winners did with the clusters, when those will come out.
Unfortunately you're not entirely correct. Most people considering Linux vs. Windows from an end-user perspective would do something on the lines of Linux stands for 'RedHat Linux' which is equated to its UI, hence Gnome. They'll say "this is how Linux looks and behaves" while looking at RH's version of the Gnome UI.
And that is because people coming from Windows don't have a high-enough resolution radar to tell the kernel apart from the UI and the producer of the distro. They look at the whole package, see the UI and call it Linux - because this is what they did with Windows all the time! Heck, even average OSX users don't single out the interface from the rest of the system!
One can only hope they'll go far enough with Linux to be able to look at the UI and call it by its proper name, be that Gnome, KDE or anything else.
So here - 2 examples of design decisions that only appealed to very few users, while obeying some arcane UI design 'rules'. Remember, UI design is an empirical science, so if more than 50% of the users don't like some setting it's the theory that is more likely to be wrong, not the users. There's no provision against making experiments in the UI, but it's quite rude to force them on people who will have close to no chance of figuring out how to change the default annoying behavior to something they can live with. And exactly this kind of 'we know better even if 95% of the users disagree' ego trip attitude was one of the big UI critiques of proprietary systems, so it would be really sad if F/OSS ends up taking the same route.
If Evolution will not work without C#, that's yet ANOTHER library you have to include with KDE - starting to get bloaty now huh?
Unless Gnome goes all-out and replaces gtk+ with gtk#, Mono (C#) will be YET ANOTHER library to include with Gnome - starting to get bloaty now huh?
Right. You were trolling.
well ... how to put it ... yahoo news had a better coverage. How's that for a hint about 'catering to the audience'?
ok, so I'm illiterate - can't tell different /. id's. sorry about putting words in your mouth. The GGP poster said that part about treating symptoms of obesity, which is funny and sad. Overeating is not an illness, it's a symptom. Obesity is a symptom, too much food and too little exercise are symptoms ... it all is about how far up the cause and effect chain you feel comfortable going - and what you want to call illness.
Did you read my post actually? I never said obesity is an hereditary illness. It is an ill state of the body, but that's not all there is - there's an ill state of mind coupled with it. By thinning out one gets a chance to break the circle and do something about it. It's quite hard (though not impossible) to make a fat person - and I mean a really fat, round one - to change his/her ways. Your advice works for losing some 40lb (the belly type), but try telling that to someone who is 150lb+ overweight. Yeah, right. Too hard - that's why surgery is so frequent, if you can afford it. Mind you, I'm not saying this (if it ever makes it out of the lab) will actually cure everyone - but for many people it could be a chance.
You, on the other hand, were talking about treating obesity symptoms (2 posts up). All that does is hook you up on medication - it cures nothing. And, if obesity is the cause for those symptoms, what would happen if this condition goes away?
(side note: here's another losing weight advice: cook your own meals. More sating than fast-food waste, so you'll eat less and use more of it. Just in case you got the impression that I was disagreeing with your post entirely. It's just that I don't think there's an one size fits all cure for obesity)
You'd have to combine it with treatment for cholesterol (and other obesity symptoms).
Do you really believe that? oh well, I guess you do, judging from the rest of the post. How sad. That's probably a clue about a medical system where often enough symptoms would get treated instead of the actual disease. Here's a hint: treat the illness and the symptoms will go away (at least, when possible - and in this case the illness is what they're aiming at).
thanks for the info! to my defense, I was plannyng to read that book ... but 'a bit' later (when I'll be sixty or so). I guess I have some of his other exquisite writings higher up on my lousy to do list right now ^_^
heh! too bad the mods decided to clue the clueless and mark you up as funny - a +5 insightful would have spawned a funny thread of replies to a funny post ^_^
no problem - with no interaction and a simple enough domain, it's trivial. Just another PDE. The devil is in the details ^_^
Given the sheer number of packages included I doubt RH would have the manpower to review every single one for stolen code. They're packaging F/OSS so it would not really be a big difference if some package authors were to commit to CVS directly in terms of responsability (and it would be for fedora, not RHEL, so the responsability should be lower). In practice setting this up is usually not that easy. And the lawyers have to be consulted anyway.
The issue seems to be RH has been dragging their feet doing this. (see the 'test code, find bugs' theme). ^_^
I agree it's a nice short email address ... as long as your gpg key is nice and long
I stand corrected :-)
... given that the previous application was submitted some 4 years before this gives rise to an interesting perspective ... say you file for some crap now and after 2-3 years snatch some new idea in the field and just amend your previous filing (ok, that's a bit overblown, but not entirely impossible). Would that make your application take precedence to any other one since the original filing date is earlier?
hmm
err ... no. pair formation is wavelength dependent - it's a threshold process. And yes, ionization (the spark) depends on absorbtion - the easier you can break electrons free the better. It's a quantum transition involved so it depends on power density (through field strength, but not wavelength here) and then on wavelength (you need to overlap the laser line with the absorbtion line of the transition).
actually, the laser is probably not focused at all - and you'd have to be talking dynamic focusing in IR and handling the full laser power (near the actual output). Plus, at that power density you'll start losing energy by air ionization long before you reach the minimum beam size. Then absorbtion skyrockets and your beam power goes the way of the dodo.
If you tune to the particular laser it's not that bad - narrow-band IR filters are actually easy. The problem is heat dissipation, but you'd already have to handle that just for air friction. Te funny thing is, if you get the thermal shield to handle enough friction heat, you might just get rid of the laser due to plasma reflectivity around the missile.
The amount of power to handle isn't 100% of the output either - the beam has diffraction (decent over km ranges), the atmosphere has absorbtion (can be fairly high in IR) and so on.
You seem to be wrong.
Their first '93 patent refers to the '87 application (not patent) as 'now abandoned'. There's no patent issued to Dana Khoyi prior to 1993 - 5,206,951 and that one was filed in '91. So any pre-'91 prior art should be ok.
as the AC said, you have to know first there's a patent in the field. This is not trivial even for older patents. Then, for any new idea one would need to go through not only awarded patents, but pending applications also. And I'm talking obscure patents filed overseas, too. Then figure out whether any vaguely related ones can be infringing. How many lawyers per programmer would that require? any way you put it, all this does is skyrocket the R&D costs of any new application.
The patent system as it is was ok for a moderate research pace with very few players in a given field. The way things are now, you have no guarantee you don't start a completely new project and by the time you're done it will be plastered with partially covering patents held by others. It might be useful to have some of the research stages published just to be able to defend against litigation companies.
well, the first patent at least is just a spec - there's no implementation attached (the filing has only 55 pages, including figures). Moreover, they use generic examples (folders, spreadsheets) about how to use the spec, not exact details (app spreadsheet101 does this and that using ObjectManager105). The bulk of the filing are function names and descriptions, field sizes and such. These are always part of the interface spec, including CORBA.
and no, it's not Kodak that developed it - they just got the patents of a dying software company ^_^
so
When will I be able to run Avalon on my Red Hat 9 machine?
no - by the time Avalon shows up o Longhorn RH9 will be ancient history and you'll be running something else on a different machine anyway ^_^